1
---
2
language: multilingual
3
license: apache-2.0
4
datasets:
5
- wikipedia
6
---
7
8
# DistilBERT base multilingual model (cased)
9
10
This model is a distilled version of the [BERT base multilingual model](bert-base-multilingual-cased). The code for the distillation process can be found
11
[here](https://github.com/huggingface/transformers/tree/master/examples/distillation). This model is cased: it does make a difference between english and English.
12
13
The model is trained on the concatenation of Wikipedia in 104 different languages listed [here](https://github.com/google-research/bert/blob/master/multilingual.md#list-of-languages).
14
The model has 6 layers, 768 dimension and 12 heads, totalizing 134M parameters (compared to 177M parameters for mBERT-base).
15
On average DistilmBERT is twice as fast as mBERT-base.
16
17
We encourage to check [BERT base multilingual model](bert-base-multilingual-cased) to know more about usage, limitations and potential biases.
18
19
| Model                        | English | Spanish | Chinese | German | Arabic  | Urdu |
20
| :---:                        | :---:   | :---:   | :---:   | :---:  | :---:   | :---:|
21
| mBERT base cased (computed)  | 82.1    | 74.6    | 69.1    | 72.3   | 66.4    | 58.5 |
22
| mBERT base uncased (reported)| 81.4    | 74.3    | 63.8    | 70.5   | 62.1    | 58.3 |
23
| DistilmBERT                  | 78.2    | 69.1    | 64.0    | 66.3   | 59.1    | 54.7 |
24
25
### BibTeX entry and citation info
26
27
```bibtex
28
@article{Sanh2019DistilBERTAD,
29
  title={DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter},
30
  author={Victor Sanh and Lysandre Debut and Julien Chaumond and Thomas Wolf},
31
  journal={ArXiv},
32
  year={2019},
33
  volume={abs/1910.01108}
34
}
35
```
36